Advertisement

Hypernatremia and moderate-to-severe hyponatremia are independent predictors of mortality in septic patients at emergency department presentation: A sub-group analysis of the need-speed trial

Published:November 05, 2020DOI:https://doi.org/10.1016/j.ejim.2020.10.003

      Highlights

      • Early risk stratification of septic patients is challenging.
      • Plasmatic sodium levels are readily available at emergency department presentation.
      • Hypernatremia and moderate/severe hyponatremia predict mortality in septic patients.
      • Dysnatremia should guide risk stratification at emergency department presentation.

      Abstract

      Study objective

      Early risk stratification of septic patients presenting to the emergency department (ED) is challenging. The aim of the study was to evaluate the prognostic role of plasmatic sodium level (PNa+) derangements at ED presentation in septic patients.

      Methods

      According to PNa+ at ED presentation patients were divided in eunatremic (136–145 mEq/L), hypernatremic (>145 mEq/L) and hyponatremic (<136 mEq/L). Hyponatremic patients were subsequently divided in mild (130–135 mEq/L), moderate (125–129 mEq/L) and severe (<125 mEq/L). 7 and 30-day mortality was evaluated according to PNa+ derangements and the degree of hyponatremia. The same analysis was then performed only in respiratory tract infection-related (RTI-r) sepsis patients.

      Results

      879 septic patients were included in this analysis, 40.3% had hyponatremia, 5.7% hypernatremia. Hypernatremia showed higher mortality rates at both endpoints compared to eunatremia and hyponatremia (p<0.0001 for both). Eunatremia and mild hyponatremia were compared vs. moderate-to-severe hyponatremia showing a significant difference in terms of 7 and 30-day survival (p = 0.004 and p = 0.007, respectively). The Cox proportional model identified as independent predictors of 7 and 30-day mortality moderate-to-severe hyponatremia (HR 4.89[2.38–10.03] and 1.79[1.07–3.01], respectively) and hypernatremia (HR 3.52[1.58–7.82] and 2.14[1.17–3.92], respectively). The same analysis was performed in patients with respiratory tract infection-related sepsis (n = 549), with similar results.

      Conclusion

      Both hypernatremia and moderate-to-severe hyponatremia at ED presentation independently predict mortality in septic patients, allowing early risk stratification and suggesting more aggressive therapeutic strategies.

      Keywords

      1. Introduction

      Early risk stratification of septic patients admitted to the Emergency Department (ED) is challenging [
      • Prasad P.A.
      • Fang M.C.
      • Abe-Jones Y.
      • Calfee C.S.
      • Matthay M.A.
      • Kangelaris K.N
      Time to recognition of sepsis in the emergency department using electronic health record data: a comparative analysis of systemic inflammatory response syndrome, sequential organ failure assessment, and quick sequential organ failure assessment.
      ,
      • Lee J.T.
      • Mikkelsen M.E.
      Risk stratification tools in sepsis: from acute physiology and chronic health evaluation to quick sequential organ failure assessment*.
      ] since signs and symptoms may be similar for patients who will have either a favorable or an adverse outcome. Among the strategies to detect patients at higher risk, some propose the clinical evaluation of tissue perfusion [
      • Hariri G.
      • Joffre J.
      • Leblanc G.
      • Bonsey M.
      • Lavillegrand J.-.R.
      • Urbina T.
      • et al.
      Narrative review: clinical assessment of peripheral tissue perfusion in septic shock.
      ,
      • Hernández G.
      • Ospina-Tascón G.A.
      • Damiani L.P.
      • Estenssoro E.
      • Dubin A.
      • Hurtado J.
      • et al.
      Effect of a resuscitation strategy targeting peripheral perfusion status vs serum lactate levels on 28-day mortality among patients with septic shock: the ANDROMEDA-SHOCK randomized clinical trial.
      ], whereas other focus on early signs of circulatory failure [
      • Gavelli F.
      • Teboul J.-.L.
      • Monnet X.
      How can CO2-derived indices guide resuscitation in critically ill patients?.
      ,
      • Wardi G.
      • Brice J.
      • Correia M.
      • Liu D.
      • Self M.
      • Tainter C.
      Demystifying lactate in the emergency department.
      ]. Again, the increase in many serum biomarkers levels, such as lactate, could alert to the need for aggressive resuscitative management. However most of these biomarkers have been studied and validated in later phases of sepsis and septic shock, notably in the intensive care unit (ICU) [
      • Pregernig A.
      • Müller M.
      • Held U.
      • Beck-Schimmer B.
      Prediction of mortality in adult patients with sepsis using six biomarkers: a systematic review and meta-analysis.
      ,
      • Castello L.M.
      • Baldrighi M.
      • Molinari L.
      • Salmi L.
      • Cantaluppi V.
      • Vaschetto R.
      • et al.
      The role of osteopontin as a diagnostic and prognostic biomarker in sepsis and septic shock.
      ,
      • Lemasle L.
      • Blet A.
      • Geven C.
      • Cherifa M.
      • Deniau B.
      • Hollinger A.
      • et al.
      Bioactive adrenomedullin, organ support therapies, and survival in the critically Ill: results from the French and European outcome registry in ICU study.
      ,
      • Baldirà J.
      • Ruiz-Rodríguez J.C.
      • Wilson D.C.
      • Ruiz-Sanmartin A.
      • Cortes A.
      • Chiscano L.
      • et al.
      Biomarkers and clinical scores to aid the identification of disease severity and intensive care requirement following activation of an in-hospital sepsis code.
      ], so that their peculiar role in the ED, where they are rarely available, has been investigated only by few [
      • Henning D.J.
      • Bhatraju P.K.
      • Johnson N.J.
      • Kosamo S.
      • Shapiro N.I.
      • Zelnick L.R.
      • et al.
      Physician judgment and circulating biomarkers predict 28-day mortality in emergency department patients.
      ,
      • Mearelli F.
      • Fiotti N.
      • Giansante C.
      • Casarsa C.
      • Orso D.
      • De Helmersen M.
      • et al.
      Derivation and validation of a biomarker-based clinical algorithm to rule out sepsis from noninfectious systemic inflammatory response syndrome at emergency department admission: a multicenter prospective study.
      ,
      • Tupchong K.
      • Chien C.
      Is procalcitonin useful in the diagnosis and treatment of acute respiratory infections in the emergency department?.
      ]. Moreover, predictive scores, such as APACHE 2 and SOFA, have been studied and validated in large multicentric studies, but the data needed to their computation are not always available during the initial evaluation performed in the ED while the qSOFA score, based only on readily available clinical data, seems to be influenced by comorbidities [
      • Lee J.T.
      • Mikkelsen M.E.
      Risk stratification tools in sepsis: from acute physiology and chronic health evaluation to quick sequential organ failure assessment*.
      ].
      On the other hand, routine tests performed at ED presentation, such as blood and urine samples, may provide useful information to detect patients who need a more aggressive treatment [
      • Engelhardt L.J.
      • Balzer F.
      • Müller M.C.
      • Grunow J.J.
      • Spies C.D.
      • Christopher K.B.
      • et al.
      Association between potassium concentrations, variability and supplementation, and in-hospital mortality in ICU patients: a retrospective analysis.
      ,
      • Dépret F.
      • Peacock W.F.
      • Liu K.D.
      • Rafique Z.
      • Rossignol P.
      • Legrand M.
      Management of hyperkalemia in the acutely ill patient.
      ]. In this regard, plasmatic sodium derangements, notably hypo- and hypernatremia, have been shown to worsen the prognosis of patients with cardiovascular, hepatic and respiratory diseases, as well as in ICU patients [
      • McCarthy K.
      • Conway R.
      • Byrne D.
      • Cournane S.
      • O'Riordan D.
      • Silke B.
      Hyponatraemia during an emergency medical admission as a marker of illness severity & case complexity.
      ,
      • Harrois A.
      • Anstey J.R.
      • Taccone F.S.
      • Udy A.A.
      • Citerio G.
      • Duranteau J.
      • et al.
      Serum sodium and intracranial pressure changes after desmopressin therapy in severe traumatic brain injury patients: a multi-centre cohort study.
      ,
      • Cárdenas A.
      • Solà E.
      • Rodríguez E.
      • Barreto R.
      • Graupera I.
      • Pavesi M.
      • et al.
      Hyponatremia influences the outcome of patients with acute-on-chronic liver failure: an analysis of the CANONIC study.
      ,
      • Müller M.
      • Schefold J.C.
      • Guignard V.
      • Exadaktylos A.K.
      • Pfortmueller C.A.
      Hyponatraemia is independently associated with in-hospital mortality in patients with pneumonia.
      ,
      • Lindner G.
      • Funk G.-.C.
      • Schwarz C.
      • Kneidinger N.
      • Kaider A.
      • Schneeweiss B.
      • et al.
      Hypernatremia in the critically ill is an independent risk factor for mortality.
      ]. Nevertheless, few studies have evaluated the prognostic role of sodium abnormalities at ED admission, specifically on septic patients, in terms of mortality.
      The aim of the present study is to evaluate whether plasma sodium concentration (PNa+) at ED admission can represent a reliable, cheap, quick and worldwide available predictor of mortality in patients with sepsis. The study derives from a subgroup analysis of the Need-Speed trial [
      • Mearelli F.
      • Fiotti N.
      • Giansante C.
      • Casarsa C.
      • Orso D.
      • De Helmersen M.
      • et al.
      Derivation and validation of a biomarker-based clinical algorithm to rule out sepsis from noninfectious systemic inflammatory response syndrome at emergency department admission: a multicenter prospective study.
      ].

      2. Materials and methods

      2.1 Patients and aims

      The aim and inclusion criteria of the Need-Speed trial have been previously published [
      • Mearelli F.
      • Fiotti N.
      • Giansante C.
      • Casarsa C.
      • Orso D.
      • De Helmersen M.
      • et al.
      Derivation and validation of a biomarker-based clinical algorithm to rule out sepsis from noninfectious systemic inflammatory response syndrome at emergency department admission: a multicenter prospective study.
      ]. Briefly, in this multicenter observational trial, consecutive adult patients admitted to the EDs of five Italian hospitals between March 2013 and March 2015, were enrolled if they met two or more criteria of systemic inflammatory response syndrome (SIRS) [

      Levy M.M., Fink M.P., Marshall J.C., Abraham E., Angus D., Cook D., et al. 2001 SCCM/ESICM/ACCP/ATS/SIS international sepsis definitions conference. Crit Care Med 2003;31:1250–6. 10.1097/01.CCM.0000050454.01978.3B.

      ].
      The aim of this subgroup analysis was to evaluate the prognostic role of sodium derangements in patients with sepsis, in terms of prediction of 7 and 30-day mortality. We, included in this analysis only patients who had been classified as “septic” according to the Sepsis-2 criteria [

      Levy M.M., Fink M.P., Marshall J.C., Abraham E., Angus D., Cook D., et al. 2001 SCCM/ESICM/ACCP/ATS/SIS international sepsis definitions conference. Crit Care Med 2003;31:1250–6. 10.1097/01.CCM.0000050454.01978.3B.

      ]. Subsequently, we performed the same analysis on patients with respiratory tract infection-related (RTI-r) sepsis.

      2.2 Data collection

      The study was approved by the Ethical Committee of each of the participating centers. At the time of inclusion, patients were informed of the study protocol and possibility was given to them to refuse participation. If clinical conditions were too serious to obtain an informed consent, the patients’ next of kin were informed of the study protocol and possibility was given to them to refuse the participation for their relatives. Subsequently, if the clinical condition had improved and if the patients were able to consent, we informed them about the study, and possibility was given to them to refuse to participate. In such a case, their data were not entered into analysis.
      At the time of enrolment demographic, hemodynamic and respiratory data were collected, together with the Glasgow Coma Scale (GCS). At the same time, arterial and peripheral venous blood samples were drawn, respectively for blood gas analysis and for cellular blood count and biochemical assays, as well as for blood culture. In addition, urine samples were collected, and patients underwent a chest X-ray, according to the clinical judgement of the treating physician.
      Patients who presented clinical, radiological or microbiological findings that suggested respiratory tract infection as the primary cause for sepsis were included in the RTI-r sepsis subgroup.

      2.3 Sodium concentration

      Among the data collected in the Need-Speed database, PNa+ at the time of ED presentation was recorded. Patients were first divided according to the presence or absence of sodium abnormalities (dysnatremic vs. non-dysnatremic, respectively) and then three different subgroups were created:
      • ü
        Eunatremia: 136≤ PNa+ ≤145 mEq/L;
      • ü
        Hyponatremia: PNa+ <136 mEq/L;
      • ü
        Hypernatremia: PNa+ >145 mEq/L.
      Subsequently, hyponatremic patients [
      • Adrogué H.J.
      • Madias N.E.
      Hyponatremia.
      ] were further divided into:
      • ü
        Mild hyponatremia: 130≤ PNa+ ≤135 mEq/L;
      • ü
        Moderate hyponatremia: 125≤ PNa+ <130 mEq/L;
      • ü
        Severe hyponatremia: PNa+ <125 mEq/L.

      2.4 Statistical analysis

      The normality of data distribution was assessed through the Kolmogorov-Smirnov test. Data are expressed as median [interquartile range] for continuous variables and as percentages for categorical variables. Comparison between groups was performed through the Mann-Whitney U test for continuous variables and through the Chi-square test for categorical variables. Multiple continuous variables were compared through the Kruskal-Wallis test, while the Bonferroni correction was used to adjust the p-value for multiple categorical variables comparisons (e.g., Chi-square test for non-2 × 2 contingency tables). Survival times among patients with different PNa+ were analyzed respectively for 7 and 30-day mortality: the log-rank test was used to identify groups with different survival probabilities and Kaplan-Meier plots were used for graphical representations. Then, variables found to be predictive of mortality with p<0.10 at the univariate analysis were introduced into a Cox proportional hazard model. Considering the distinct difference between patients with moderate/severe hyponatremia and patients with mild PNa+ derangement/eunatremia, a further comparison was then performed by constituting two non-pre-specified groups: “moderate-to-severe hyponatremia” group vs. “eunatremia+mild hyponatremia”. Similarly, a Cox proportional hazard model was built to identify independent predictors of mortality at 7 and 30-day in the overall population, including as categorical variables “eunatremia+mild hyponatremia”, and “moderate-to-severe hyponatremia” or hypernatremia.
      Statistical significance was set at two-tailed p<0.05. The statistical analysis was performed with MedCalc 19.3.1 software (Mariakerke, Belgium).

      3. Results

      3.1 Patients

      Among the 1132 patients included in the primary analysis of the Need-Speed trial [
      • Mearelli F.
      • Fiotti N.
      • Giansante C.
      • Casarsa C.
      • Orso D.
      • De Helmersen M.
      • et al.
      Derivation and validation of a biomarker-based clinical algorithm to rule out sepsis from noninfectious systemic inflammatory response syndrome at emergency department admission: a multicenter prospective study.
      ], 890 patients with a definitive diagnosis of sepsis were considered eligible. Eleven of them were excluded because PNa+ was not available: the analysis was therefore performed on 879 patients. The flowchart of patients’ selection is reported in Supplementary Material S1. The main characteristics of the overall sample are presented in Table 1.
      Table 1Patient characteristics.
      Patient characteristics (n = 879)
      Age (years)80 [72 - 87]
      Sex (M/F)475 / 404
      HR (bpm)100 [90 - 110]
      RR (breaths per minute)24 [20 - 28]
      MAP (mmHg)87 [77 - 97]
      SpO2 (%)94 [92 - 96]
      Hb (g/dL)12.2 [10.8 - 13.4]
      Ht (%)37.7 [33.0 - 41.1]
      WBCs (x103/mm3)12.90 [9.36 - 17.11]
      PLTs (x103/mm3)222 [158 - 298]
      PNa+ (mEq/L)136 [133 - 139]
      PK+ (mEq/L)3.96 [3.57 - 4.40]
      Creatinine (mg/dL)1.08 [0.83 - 1.67]
      Total bilirubin (mg/dL)0.91 [0.66 - 1.43]
      INR1.19 [1.10 - 1.36]
      Arterial pH7.45 [7.40 - 7.48]
      PaCO2 (mmHg)34.1 [30.0 - 39.7]
      PaO2 (mmHg)69.5 [59.3 - 82.4]
      PHCO3 (mmol/L)23.7 [21.1 - 26.3]
      PaO2/FiO2286 [231 - 346]
      Lactate (mmol/L)1.5 [1.1 - 2.2]
      CRP (mg/dL)10.21 [3.42 - 18.83]
      GCS15 [15 - 15]
      PNa+ Hypernatremic (mEq/L) (n = 50)150 [148 - 155]
      PNa+ Eunatremic (mEq/L) (n = 451)138 [136 - 140]
      PNa+ Hyponatremic (mEq/L) (n = 378)132 [129 - 133]
      PNa+ Mild Hyponatremic (n = 269)

      PNa+ Moderate Hyponatremic (n = 89)

      PNa+ Severe Hyponatremic (n = 20)
      133 [131 - 134]

      128 [127 - 129]

      122 [120 - 124]
      CRP: C-reactive protein; GCS: Glasgow Coma Scale; Hb: hemoglobin; Ht: hematocrit; HR: heart rate; INR: international normalized ratio; MAP: mean arterial pressure; PaCO2: arterial partial pressure of carbon dioxide; PaO2: arterial partial pressure of oxygen; PaO2/FiO2: ratio between partial pressure of oxygen and fractional inspired oxygen; PLTs: platelets; RR: respiratory rate; SpO2: peripheral oxygen saturation; WBCs: white blood cells; PHCO3: plasma bicarbonate concentration; PK+: plasma potassium concentration; PNa+: plasma sodium concentration.

      3.2 Prognostic value of PNa+ in the overall sepsis population

      The prognostic value of PNa+ was investigated in the septic population according to 7-day (n = 85, 9.7%) and 30-day (n = 173, 19.7%) mortality. At both endpoints PNa+ was found to be higher in non-survivors compared to survivors. The main demographic, physical and laboratory characteristics of survivors and non-survivors at each endpoint, as well as the results of the univariate analyses, are reported in Supplementary Material S2. Half of the enrolled patients was admitted to the ED with a sodium derangement (48.7%), either in terms of hyponatremia (n = 378, 43.0%) or hypernatremia (n = 50, 5.7%). Compared to eunatremic patients, the dysnatremic ones had significantly higher mortality rates at 7 days, but the difference did not reach statistical significance at 30 days (11.9% vs. 7.5% p = 0.038 and 22.2% vs.17.3% p = 0.082, respectively) (Table 2).
      Table 2Comparison of mortality rates at 7 and 30 days according to different degrees of sodium derangements.
      7-day mortality30-day mortality
      Non-survivors%Non-survivors%
      Dysnatremia5111.9%9522.2%
      Eunatremia347.5%
      statistically significant vs. Dysnatremia on 7-day mortality.
      7817.3%
      Hyponatremia338.7%
      statistically significant vs. Hypernatremia on 7-day mortality.
      7018.5%
      statistically significant vs. Hypernatremia on 30-day mortality.
      Eunatremia347.5%
      statistically significant vs. Hypernatremia on 7-day mortality.
      7817.3%
      statistically significant vs. Hypernatremia on 30-day mortality.
      Hypernatremia1836.0%2550.0%
      Eunatremia347.5%7817.3%
      Mild hyponatremia165.9%4014.9%
      Moderate hyponatremia1415.7%
      statistically significant vs. Mild hyponatremia on 7-day mortality.
      2528.1%
      statistically significant vs. Mild hyponatremia on 30-day mortality.
      Severe hyponatremia315.0%525.0%
      Eunatremia +

      mild hyponatremia
      506.9%11816.4%
      Moderate-to-severe

      hyponatremia
      1715.6%
      statistically significant vs. Eunatremia + mild hyponatremia on 7-day mortality.
      3027.5%
      statistically significant vs. Eunatremia + mild hyponatremia on 30-day mortality.
      The Chi-square test was used to compare mortality rates of each row (notably Dysnatremia vs. Eunatremia, Hyponatremia vs. Eunatremia vs. Hypernatremia, Eunatremia vs. Mild hyponatremia vs. Moderate hyponatremia vs. Severe hyponatremia and Eunatremia+mild hyponatremia vs. Moderate-to-severe hyponatremia), with the Bonferroni correction for multiple comparisons, as appropriate. Percentages refer to the overall population.
      a statistically significant vs. Dysnatremia on 7-day mortality.
      b statistically significant vs. Hypernatremia on 7-day mortality.
      c statistically significant vs. Hypernatremia on 30-day mortality.
      d statistically significant vs. Mild hyponatremia on 7-day mortality.
      e statistically significant vs. Mild hyponatremia on 30-day mortality.
      f statistically significant vs. Eunatremia + mild hyponatremia on 7-day mortality.
      g statistically significant vs. Eunatremia + mild hyponatremia on 30-day mortality.
      When dysnatremic patients were further divided according to the nature of the sodium derangement, hypernatremia was associated with significantly higher mortality rates at both endpoints compared to eunatremia and hyponatremia (p<0.0001 for both). No significant difference was observed between eunatremic and hyponatremic patients (Table 2).
      In order to detect a possible confounding role of hyperglycemia-related hyponatremia, in 123 patients with plasma glucose concentration ≥ 200 mg/dL we adjusted PNa+ values through the Hillier formula [
      • Hillier T.A.
      • Abbott R.D.
      • Barrett E.J.
      Hyponatremia: evaluating the correction factor for hyperglycemia.
      ]. With such operation, only 48 patients (5.5%) were classified differently: three patients shifted from eunatremia to hypernatremia and the remaining ones from hyponatremia to eunatremia, with no significant changes in mortality rates among groups.

      3.3 Hyponatremia in the overall sepsis population

      As pre-specified, a subsequent analysis was performed by excluding hypernatremic patients and by dividing the hyponatremic ones in three groups according to the severity of PNa+ derangement: mild (n = 269, 71.2%), moderate (n = 89, 23.5%) and severe (n = 20, 5.3%) hyponatremia. The comparison performed among these three groups and patients with eunatremia highlighted a significant difference in mortality rates at 7 and 30 days. However, when the p-value was adjusted for multiple comparisons, only moderate hyponatremia vs. mild hyponatremia, both at 7 and 30 days, confirmed a statistical significance (respectively p = 0.008 and p = 0.008) (Table 2).
      Moreover, the “moderate-to-severe hyponatremia” group showed a significantly higher mortality at each endpoint compared to the “eunatremia+mild hyponatremia” one (p = 0.004 and p = 0.007 at 7 and 30 days, respectively) (Table 2). When PNa+ values were corrected for glucose levels, only 13 hyponatremic patients (3.4%) were classified differently: two patients shifted from severe to moderate and eleven from moderate to mild hyponatremia, again with no changes in mortality among groups.

      3.4 Survival analysis in the overall sepsis population

      A survival analysis was performed by dividing patients in the above-mentioned categories. The Kaplan-Meier curves that graphically show comparisons between eunatremia vs. dysnatremia and between “eunatremia+mild hyponatremia” vs. “moderate-to-severe hyponatremia” are presented in Fig. 1 and Supplementary Material S3, respectively. When hypernatremia was added to the latter analysis, a significant difference was observed in terms of survival between the three groups (Fig. 2).
      Fig. 1
      Fig. 1Kaplan-Meier curve for the comparison of 7 and 30-day mortality between eunatremic vs. dysnatremic patients. As indicated by the p-values for the log-rank tests, statistical significance was reached only at 7 days.
      Fig. 2
      Fig. 2Kaplan-Meier curve for the comparison of 7 and 30-day mortality between patients presenting eunatremia + mild hyponatremia vs. moderate-to-severe hyponatremia vs. hypernatremia. As indicated by the p-values for the log-rank tests, statistical significance was reached at both endpoints.
      A Cox proportional hazard model was built to identify independent predictors of mortality at 7 and 30-day in the overall population. Since both hyponatremia and hypernatremia were associated to higher mortality rates (Supplementary Material S4), PNa+ was not embedded in the model as a continuous variable. On the contrary, the model included a categorical variable i.e. “eunatremia+mild hyponatremia”, and “moderate-to-severe hyponatremia” or hypernatremia. These categories were selected based on our previous analysis. The Cox proportional hazard identified as independent predictors of both 7 and 30-day mortality “moderate-to-severe hyponatremia” (HR 4.89 [2.38–10.03] and 1.79 [1.07–3.01], respectively) and hypernatremia (HR 3.52 [1.58–7.82] and 2.14 [1.17–3.92], respectively), together with age and plasma lactate (Table 3). Amongst the other variables significantly associated to mortality at the univariate analysis, at the multivariate one the GCS did not emerge as independent predictor of mortality either at 7 or at 30 days. Of note, sex differences were not associated either with different distribution of sodium derangements or with mortality rates at both endpoints.
      Table 37 and 30-day independent predictors of mortality in the overall population.
      7-day mortality30-day mortality
      HR95% CIp-valueHR95% CIp-value
      Moderate-to-severe

      hyponatremia
      4.892.38 - 10.03<0.0001Moderate-to-severe

      hyponatremia
      1.791.07 - 3.010.029
      Hypernatremia3.521.58 - 7.820.002Hypernatremia2.141.17 - 3.920.014
      Plasma lactate1.151.05- 1.260.002Plasma lactate1.181.09- 1.28<0.0001
      Age1.041.01 - 1.070.013Age1.041.02 - 1.060.001
      Heart rate1.021.00 - 1.030.046PaO2/FiO20.990.99 - 0.990.001
      Arterial pH0.010.00 - 0.430.015SpO20.960.92 - 0.990.011
      Hb0.880.79 - 0.970.012
      The Table shows the HR resulted from the multivariate analysis conducted through Cox proportional hazard regression models.
      CI: confidence interval; Hb: hemoglobin; HR: hazard ratio; PaO2/FiO2: ratio of arterial oxygen partial pressure over the oxygen inspired fraction; SpO2: peripheral oxygen saturation.

      3.5 Respiratory tract infection related (RTI-r) sepsis population

      Five hundred forty-nine patients (62.5%), out of the whole group of 879 patients, presented clinical, radiological or microbiological findings suggestive of RTI-r sepsis. The characteristics of patients with RTI-r sepsis and with sepsis from other sites (non-RTI-r sepsis) are reported in Supplementary Material S5. Mortality rates were significantly higher in RTI-r sepsis compared to non-RTI-r sepsis patients both at 7 and 30 days (log-rank test: p = 0.001 and p<0.0001, respectively) (Supplementary Material S6).

      3.6 Sodium derangements in RTI-r sepsis and non-RTI-r sepsis population

      Forty-seven percent of RTI-r sepsis patients presented a sodium derangement: 39.5% were hyponatremic and 7.1% were hypernatremic. Furthermore, among hyponatremic patients, the derangement was mild in 70.5%, moderate in 23.5% and severe in 6.0%. Compared to the non-RTI-r sepsis group, no significant difference was observed in the prevalence of sodium derangements (p = 0.132). When the type of sodium derangement was compared among the two populations, hypernatremia emerged to be more frequent in RTI-r sepsis and hyponatremia more frequent in non-RTI-r sepsis patients (Supplementary Material S7).

      3.7 Prognostic value of PNa+ and survival analysis in the RTI-r sepsis population

      When the characteristics of survivors and non-survivors were compared in RTI-r sepsis patients, PNa+ was found to be significantly higher in non-survivors both at 7 (p = 0.045) and 30 days (p = 0.028) (Supplementary Material S8). As for the overall sepsis population, mortality rates were significantly higher in dysnatremic patients both at 7 and 30 days. When dysnatremia was divided in hyponatremia and hypernatremia, mortality was significantly higher in hypernatremic patients at both endpoints (Supplementary Material S9). Again, we excluded hypernatremic patients and divided the remaining subjects in four groups (notably, eunatremia, mild, moderate and severe hyponatremia): only the mortality rate between moderate hyponatremia and eunatremia was significantly different (p = 0.008). When such patients were combined in the same previous two larger groups, mortality rates were significantly higher in those with “moderate-to-severe hyponatremia” compared to those with “eunatremia+mild hyponatremia” at both endpoints (Supplementary Material S9).
      Then, survival analyses were performed according to the same combination of categories as for the general population (Supplementary Material S10). The Cox proportional hazard model subsequently built, identified both “moderate-to-severe hyponatremia” and hypernatremia as independent predictors of both 7 and 30-day mortality (Supplementary Material S11).

      3.8 Prognostic value of pNa+ in the non-RTI-r sepsis population

      On the contrary, no differences were found in pNa+ between survivors and non-survivors in the non-RTI-r sepsis population. Furthermore, when these patients were divided in the same categories listed in the previous paragraph, sodium derangements of any kind (either hyponatremia or hypernatremia) or degree (mild, moderate or severe hyponatremia, or even combined eunatremia+mild hyponatremia and moderate-to-severe hyponatremia) were not associated with higher mortality rates both at 7 and 30 days.

      4. Discussion

      This subgroup analysis of the Need-Speed trial shows that in septic patients, both hypernatremia and moderate-to-severe hyponatremia at ED presentation are independent predictors of mortality at both 7 and 30 days. When considering only patients with RTI-r sepsis, similar results were observed.
      Detecting which septic patients are at higher risk of death from the first hours is a real need for the emergency and critical care physician. As a matter of fact, even the existing guidelines do not provide elements for a different management according to the a priori patient's risk [
      • Rhodes A.
      • Evans L.E.
      • Alhazzani W.
      • Levy M.M.
      • Antonelli M.
      • Ferrer R.
      • et al.
      Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016.
      ]. Therefore, the same treatment is suggested for patients who will likely have a less severe clinical course as well as for patients who will require a more aggressive therapeutic management. To overcome this issue and stratify the risk of death, different scores have been developed. However, they require the collection of many information and have been validated only in the ICU [
      • Gall J.-.R.L.
      • Lemeshow S.
      • Saulnier F.
      A new simplified acute physiology score (SAPS II) based on a European/North American Multicenter Study.
      ,
      • Vincent J.L.
      • Moreno R.
      • Takala J.
      • Willatts S.
      • De Mendonça A.
      • Bruining H.
      • et al.
      The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the working group on sepsis-related problems of the European Society of Intensive Care Medicine.
      ,
      • Knaus W.A.
      • Wagner D.P.
      • Draper E.A.
      • Zimmerman J.E.
      • Bergner M.
      • Bastos P.G.
      • et al.
      The APACHE III prognostic system: risk prediction of hospital mortality for critically III hospitalized adults.
      ]. As the first medical contact of septic patients is at ED presentation, simple and rapidly available information are needed to detect patients at higher risk.
      Beyond serum biomarkers measurement, data from routinely performed biochemical analyses have been proposed for this purpose, with encouraging results. Semeraro et al.[
      • Semeraro F.
      • Ammollo C.T.
      • Caironi P.
      • Masson S.
      • Latini R.
      • Panigada M.
      • et al.
      D-dimer corrected for thrombin and plasmin generation is a strong predictor of mortality in patients with sepsis.
      ]. have shown that D-Dimer could predict mortality in septic patients at ED admission, provided that its value is corrected for thrombin and fibrin generation. The prognostic role of lactate has been recognized [
      • Bakker J.
      • Nijsten M.W.
      • Jansen T.C.
      Clinical use of lactate monitoring in critically ill patients.
      ], although not the initial measurement, but rather its variation over time has the highest prognostic value [
      • Mitra B.
      • Roman C.
      • Charters K.E.
      • O'Reilly G.
      • Gantner D.
      • Cameron P.A
      Lactate, bicarbonate and anion gap for evaluation of patients presenting with sepsis to the emergency department: a prospective cohort study.
      ,
      • Gattinoni L.
      • Vasques F.
      • Camporota L.
      • Meessen J.
      • Romitti F.
      • Pasticci I.
      • et al.
      Understanding lactatemia in human sepsis. Potential impact for early management.
      ]. Again, the venous-arterial difference in the carbon dioxide partial pressure (PCO2-gap) has been shown to predict mortality in patients with sepsis at ED admission [
      • Shetty A.
      • Sparenberg S.
      • Adams K.
      • Selvedran S.
      • Tang B.
      • Hanna K.
      • et al.
      Arterial to end-tidal carbon dioxide tension difference (CO2 gap) as a prognostic marker for adverse outcomes in emergency department patients presenting with suspected sepsis.
      ], but its computation may not be practical in the ED. Thus, stratifying patients’ prognosis simply on the first measurement of plasmatic sodium, without the need of further measurements or score calculation, may be very helpful in detecting patients who will need more intensive treatment.
      PNa+ derangements are frequently encountered in the general hospitalized population [
      • Reynolds R.M.
      • Padfield P.L.
      • Seckl J.R.
      Disorders of sodium balance.
      ,
      • Castello L.M.
      • Baldrighi M.
      • Panizza A.
      • Bartoli E.
      • Avanzi G.C.
      Efficacy and safety of two different tolvaptan doses in the treatment of hyponatremia in the emergency department.
      ], and it is observed in up to 5% and 35% of patients for hypernatremia and hyponatremia, respectively [
      • Molaschi M.
      • Ponzetto M.
      • Massaia M.
      • Villa L.
      • Scarafiotti C.
      • Ferrario E.
      Hypernatremic dehydration in the elderly on admission to hospital.
      ,]. However, in critically ill patients their incidence is estimated to be higher, with direct impact on patients’ outcome [
      • Overgaard-Steensen C.
      • Ring T.
      Clinical review: practical approach to hyponatraemia and hypernatraemia in critically ill patients.
      ,
      • Rosner M.H.
      • Ronco C.
      Dysnatremias in the intensive care unit.
      ]. As a matter of fact, in ICU patients, both the presence and the degree of hypernatremia are indicative of the severity of the underlying disease, and are related to increased mortality [
      • Tsipotis E.
      • Price L.L.
      • Jaber B.L.
      • Madias N.E.
      Hospital-associated hypernatremia spectrum and clinical outcomes in an unselected cohort.
      ,
      • Polderman K.H.
      • Schreuder W.O.
      • Strack van Schijndel R.J.
      • Thijs L.G.
      Hypernatremia in the intensive care unit: an indicator of quality of care?.
      ] and hyponatremia is associated to an increased ICU and in-hospital length of stay. When PNa+ abnormalities are severe, the risk of death reaches up to 40% [
      • Pasantes-Morales H.
      • Cruz-Rangel S.
      Brain volume regulation: osmolytes and aquaporin perspectives.
      ].
      To this regard, our analysis confirms the prognostic role of both hypernatremia and moderate-to-severe hyponatremia in the specific population of septic patients, even though we cannot establish whether such sodium alterations are causative or rather associated to sepsis development. However, compared to the previous studies conducted in the ICU, we show that such predictive power can be detected very early, during the first medical contact in the ED, thus providing an immediate hint for risk stratification.
      In the first part of our study, we aimed at analyzing the overall septic population of the Need-Speed cohort, regardless of the primary site of infection. Interestingly, we confirm that hypernatremia is less common than hyponatremia. However, the mortality rates of this group of septic patients were 36% and 50% at 7 and 30 days, respectively, which is slightly higher than reported in literature [
      • Tsipotis E.
      • Price L.L.
      • Jaber B.L.
      • Madias N.E.
      Hospital-associated hypernatremia spectrum and clinical outcomes in an unselected cohort.
      ,
      • Polderman K.H.
      • Schreuder W.O.
      • Strack van Schijndel R.J.
      • Thijs L.G.
      Hypernatremia in the intensive care unit: an indicator of quality of care?.
      ]. As hypernatremia is more often related to a free water deficit issue rather that sodium homeostasis, in patients admitted to the ED with sepsis, thus presenting both absolute and relative hypovolemia [
      • Scheeren T.W.L.
      • Bakker J.
      • De Backer D.
      • Annane D.
      • Asfar P.
      • Boerma E.C.
      • et al.
      Current use of vasopressors in septic shock.
      ], failure of fluid resuscitation efforts becomes more likely [
      • Overgaard-Steensen C.
      • Ring T.
      Clinical review: practical approach to hyponatraemia and hypernatraemia in critically ill patients.
      ,
      • Beurton A.
      • Teboul J.-.L.
      • Gavelli F.
      • Gonzalez F.A.
      • Girotto V.
      • Galarza L.
      • et al.
      The effects of passive leg raising may be detected by the plethysmographic oxygen saturation signal in critically ill patients.
      ,
      • Ni H.-.B.
      • Hu X.-.X.
      • Huang X.-.F.
      • Liu K.-.Q.
      • Yu C.-.B.
      • Wang X.-.M.
      • et al.
      Risk factors and outcomes in patients with hypernatremia and sepsis.
      ]. In this sub-group, due to the relatively small number of patients, we could not perform a further analysis on mortality according to the severity of hypernatremia. However, at the multivariate analysis, hypernatremia emerged as an independent risk factor of mortality with a hazard ratio of 3.5 (Table 3).
      On the other hand, patients with hyponatremia showed a dichotomous behavior. In fact, those with mild hyponatremia had no increased risk of dying compared to eunatremic patients. This is not surprising, as mild hyponatremia is the most common type of electrolyte disturbance in hospitalized patients and, in many cases, it is not related to a real reduction in the total amount of sodium, but it is rather the manifestation of a chronic medical condition [
      • Adrogué H.J.
      • Madias N.E.
      Hyponatremia.
      ,
      • Reynolds R.M.
      • Padfield P.L.
      • Seckl J.R.
      Disorders of sodium balance.
      ,
      • Baldrighi M.
      • Sainaghi P.P.
      • Bellan M.
      • Bartoli E.
      • Castello L.M.
      Hyperglycemic hyperosmolar state: a pragmatic approach to properly manage sodium derangements.
      ]. Although it is often asymptomatic, it has been suggested that even borderline derangements of sodium concentration may worsen patients’ prognosis [
      • Darmon M.
      • Diconne E.
      • Souweine B.
      • Ruckly S.
      • Adrie C.
      • Azoulay E.
      • et al.
      Prognostic consequences of borderline dysnatremia: pay attention to minimal serum sodium change.
      ]. On the opposite, in our cohort the two more severe degrees of hyponatremia were similarly associated to an increased mortality. To highlight this difference, we combined moderate and severe hyponatremia and demonstrated that such derangements are independently responsible for a 5-fold increase in mortality at 7 days in septic patients, and that such prediction can be observed in the ED at the first medical contact (Table 3).
      When patients were divided according to the source of infection, we observed different results in the RTI-r sepsis group and in the non-RTI-r sepsis group: in fact, sodium derangements were strongly associated with mortality in patients with sepsis of respiratory tract origin, while no association was found in sepsis from other sources. This was not surprising, since RTIs are often accompanied by sodium derangements, especially hyponatremia, which independently predicts a worse outcome [
      • Müller M.
      • Schefold J.C.
      • Guignard V.
      • Exadaktylos A.K.
      • Pfortmueller C.A.
      Hyponatraemia is independently associated with in-hospital mortality in patients with pneumonia.
      ] and is therefore included in many prognostic scores, one for all the Pneumonia Severity Index or PSI [
      • Valencia M.
      • Badia J.R.
      • Cavalcanti M.
      • Ferrer M.
      • Agustí C.
      • Angrill J.
      • et al.
      Pneumonia severity index class v patients with community-acquired pneumonia: characteristics, outcomes, and value of severity scores.
      ]. In this regard, our results confirm data from the literature, as the 7-day risk of mortality was increased by 5 times in moderate-to-severe hyponatremic patients. Moreover, we demonstrated that also hypernatremia is independently responsible for a similar decrease in survival at 7 days. Again, such information can be derived at ED admission: since mortality of RTI-r sepsis is higher compared to the one from other sources [
      • Tillmann B.
      • Wunsch H.
      Epidemiology and outcomes.
      ], such an early stratification may prompt aggressive therapeutic management, eventually leading to a more favorable outcome.

      4.1 Limitations

      First, this is a sub-group analysis of prospectively enrolled patients for another trial [
      • Mearelli F.
      • Fiotti N.
      • Giansante C.
      • Casarsa C.
      • Orso D.
      • De Helmersen M.
      • et al.
      Derivation and validation of a biomarker-based clinical algorithm to rule out sepsis from noninfectious systemic inflammatory response syndrome at emergency department admission: a multicenter prospective study.
      ]. Thus, our outcomes were not pre-specified at the time of the study design. Similarly, patients’ inclusion was already performed. Nevertheless, as only 1% of such septic patients had to be excluded from our study, we judged the selection bias in our subgroup analysis as “low”. Second, we used the Sepsis-2 criteria for patient's inclusion, which were published long time ago. However, the Need-Speed Trial was conducted between 2013 and 2015, while the latest Sepsis-3 criteria have appeared only in 2016 [
      • Singer M.
      • Deutschman C.S.
      • Seymour C.W.
      • Shankar-Hari M.
      • Annane D.
      • Bauer M.
      • et al.
      The third international consensus definitions for sepsis and septic shock (Sepsis-3).
      ]. Third, we could not retrieve information related to patients’ chronic treatment, or the presence of comorbidities which could have acted as possible confounders. Hence, we could not evaluate whether sodium derangements were influenced by drugs interfering with its plasmatic levels or chronic medical conditions. However, it is unlikely that such information would have affected the results of our study, as the early prognostic role of sodium derangement seems quite strong. Fourth, we recorded baseline value of plasmatic sodium, and we could not evaluate the kinetics of sodium correction during hospitalization. However, the aim of the present analysis was to identify an easy and immediate factor able to could stratify the mortality risk of patients with sepsis, at the first medical contact in the ED.

      5. Conclusions

      In septic patients the presence of either hypernatremia or moderate-to-severe hyponatremia at ED presentation are independent predictors of 7 and 30-day mortality. This information provides an immediate risk stratification and should promote more aggressive therapeutic management of such patients.

      Authors’ contribution

      LMC, FG and GCA conceived and designed the study.
      MBa, LS, FP, MBe, FM, NF, GR, SD, EL, LMM and GB collected the data.
      LMC, FG and MBa analyzed and interpreted the data.
      LMC, FG, MBa drafted the report and all authors contributed to review it.
      All authors approved the final version.

      Declaration of Competing Interest

      All the authors state they do not have any conflict of interest to declare.

      Appendix. Supplementary materials

      References

        • Prasad P.A.
        • Fang M.C.
        • Abe-Jones Y.
        • Calfee C.S.
        • Matthay M.A.
        • Kangelaris K.N
        Time to recognition of sepsis in the emergency department using electronic health record data: a comparative analysis of systemic inflammatory response syndrome, sequential organ failure assessment, and quick sequential organ failure assessment.
        Crit Care Med. 2020; 48: 200-209https://doi.org/10.1097/CCM.0000000000004132
        • Lee J.T.
        • Mikkelsen M.E.
        Risk stratification tools in sepsis: from acute physiology and chronic health evaluation to quick sequential organ failure assessment*.
        Crit Care Med. 2019; 47: 1159-1161https://doi.org/10.1097/CCM.0000000000003859
        • Hariri G.
        • Joffre J.
        • Leblanc G.
        • Bonsey M.
        • Lavillegrand J.-.R.
        • Urbina T.
        • et al.
        Narrative review: clinical assessment of peripheral tissue perfusion in septic shock.
        Ann Intensive Care. 2019; 9: 37https://doi.org/10.1186/s13613-019-0511-1
        • Hernández G.
        • Ospina-Tascón G.A.
        • Damiani L.P.
        • Estenssoro E.
        • Dubin A.
        • Hurtado J.
        • et al.
        Effect of a resuscitation strategy targeting peripheral perfusion status vs serum lactate levels on 28-day mortality among patients with septic shock: the ANDROMEDA-SHOCK randomized clinical trial.
        JAMA. 2019; 321: 654-664https://doi.org/10.1001/jama.2019.0071
        • Gavelli F.
        • Teboul J.-.L.
        • Monnet X.
        How can CO2-derived indices guide resuscitation in critically ill patients?.
        J Thorac Dis. 2019; 11: S1528-S1537https://doi.org/10.21037/jtd.2019.07.10
        • Wardi G.
        • Brice J.
        • Correia M.
        • Liu D.
        • Self M.
        • Tainter C.
        Demystifying lactate in the emergency department.
        Ann Emerg Med. 2020; 75: 287-298https://doi.org/10.1016/j.annemergmed.2019.06.027
        • Pregernig A.
        • Müller M.
        • Held U.
        • Beck-Schimmer B.
        Prediction of mortality in adult patients with sepsis using six biomarkers: a systematic review and meta-analysis.
        Ann Intensive Care. 2019; 9: 125https://doi.org/10.1186/s13613-019-0600-1
        • Castello L.M.
        • Baldrighi M.
        • Molinari L.
        • Salmi L.
        • Cantaluppi V.
        • Vaschetto R.
        • et al.
        The role of osteopontin as a diagnostic and prognostic biomarker in sepsis and septic shock.
        Cells. 2019; 8https://doi.org/10.3390/cells8020174
        • Lemasle L.
        • Blet A.
        • Geven C.
        • Cherifa M.
        • Deniau B.
        • Hollinger A.
        • et al.
        Bioactive adrenomedullin, organ support therapies, and survival in the critically Ill: results from the French and European outcome registry in ICU study.
        Crit Care Med. 2020; 48: 49-55https://doi.org/10.1097/CCM.0000000000004044
        • Baldirà J.
        • Ruiz-Rodríguez J.C.
        • Wilson D.C.
        • Ruiz-Sanmartin A.
        • Cortes A.
        • Chiscano L.
        • et al.
        Biomarkers and clinical scores to aid the identification of disease severity and intensive care requirement following activation of an in-hospital sepsis code.
        Ann Intensive Care. 2020; 10: 7https://doi.org/10.1186/s13613-020-0625-5
        • Henning D.J.
        • Bhatraju P.K.
        • Johnson N.J.
        • Kosamo S.
        • Shapiro N.I.
        • Zelnick L.R.
        • et al.
        Physician judgment and circulating biomarkers predict 28-day mortality in emergency department patients.
        Crit Care Med. 2019; 47: 1513-1521https://doi.org/10.1097/CCM.0000000000003899
        • Mearelli F.
        • Fiotti N.
        • Giansante C.
        • Casarsa C.
        • Orso D.
        • De Helmersen M.
        • et al.
        Derivation and validation of a biomarker-based clinical algorithm to rule out sepsis from noninfectious systemic inflammatory response syndrome at emergency department admission: a multicenter prospective study.
        Crit Care Med. 2018; 46: 1421-1429https://doi.org/10.1097/CCM.0000000000003206
        • Tupchong K.
        • Chien C.
        Is procalcitonin useful in the diagnosis and treatment of acute respiratory infections in the emergency department?.
        Ann Emerg Med. 2018; 72: 57-58https://doi.org/10.1016/j.annemergmed.2018.01.040
        • Engelhardt L.J.
        • Balzer F.
        • Müller M.C.
        • Grunow J.J.
        • Spies C.D.
        • Christopher K.B.
        • et al.
        Association between potassium concentrations, variability and supplementation, and in-hospital mortality in ICU patients: a retrospective analysis.
        Ann Intensive Care. 2019; 9: 100https://doi.org/10.1186/s13613-019-0573-0
        • Dépret F.
        • Peacock W.F.
        • Liu K.D.
        • Rafique Z.
        • Rossignol P.
        • Legrand M.
        Management of hyperkalemia in the acutely ill patient.
        Ann Intensive Care. 2019; 9: 32https://doi.org/10.1186/s13613-019-0509-8
        • McCarthy K.
        • Conway R.
        • Byrne D.
        • Cournane S.
        • O'Riordan D.
        • Silke B.
        Hyponatraemia during an emergency medical admission as a marker of illness severity & case complexity.
        Eur J Intern Med. 2019; 59: 60-64https://doi.org/10.1016/j.ejim.2018.08.002
        • Harrois A.
        • Anstey J.R.
        • Taccone F.S.
        • Udy A.A.
        • Citerio G.
        • Duranteau J.
        • et al.
        Serum sodium and intracranial pressure changes after desmopressin therapy in severe traumatic brain injury patients: a multi-centre cohort study.
        Ann Intensive Care. 2019; 9: 99https://doi.org/10.1186/s13613-019-0574-z
        • Cárdenas A.
        • Solà E.
        • Rodríguez E.
        • Barreto R.
        • Graupera I.
        • Pavesi M.
        • et al.
        Hyponatremia influences the outcome of patients with acute-on-chronic liver failure: an analysis of the CANONIC study.
        Crit Care. 2014; 18: 700https://doi.org/10.1186/s13054-014-0700-0
        • Müller M.
        • Schefold J.C.
        • Guignard V.
        • Exadaktylos A.K.
        • Pfortmueller C.A.
        Hyponatraemia is independently associated with in-hospital mortality in patients with pneumonia.
        Eur J Intern Med. 2018; 54: 46-52https://doi.org/10.1016/j.ejim.2018.04.008
        • Lindner G.
        • Funk G.-.C.
        • Schwarz C.
        • Kneidinger N.
        • Kaider A.
        • Schneeweiss B.
        • et al.
        Hypernatremia in the critically ill is an independent risk factor for mortality.
        Am J Kidney Dis. 2007; 50: 952-957https://doi.org/10.1053/j.ajkd.2007.08.016
      1. Levy M.M., Fink M.P., Marshall J.C., Abraham E., Angus D., Cook D., et al. 2001 SCCM/ESICM/ACCP/ATS/SIS international sepsis definitions conference. Crit Care Med 2003;31:1250–6. 10.1097/01.CCM.0000050454.01978.3B.

        • Adrogué H.J.
        • Madias N.E.
        Hyponatremia.
        N Engl J Med. 2000; 342: 1581-1589https://doi.org/10.1056/NEJM200005253422107
        • Hillier T.A.
        • Abbott R.D.
        • Barrett E.J.
        Hyponatremia: evaluating the correction factor for hyperglycemia.
        Am J Med. 1999; 106: 399-403https://doi.org/10.1016/s0002-9343(99)00055-8
        • Rhodes A.
        • Evans L.E.
        • Alhazzani W.
        • Levy M.M.
        • Antonelli M.
        • Ferrer R.
        • et al.
        Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016.
        Intensive Care Med. 2017; 43: 304-377https://doi.org/10.1007/s00134-017-4683-6
        • Gall J.-.R.L.
        • Lemeshow S.
        • Saulnier F.
        A new simplified acute physiology score (SAPS II) based on a European/North American Multicenter Study.
        JAMA. 1993; 270: 2957-2963https://doi.org/10.1001/jama.1993.03510240069035
        • Vincent J.L.
        • Moreno R.
        • Takala J.
        • Willatts S.
        • De Mendonça A.
        • Bruining H.
        • et al.
        The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the working group on sepsis-related problems of the European Society of Intensive Care Medicine.
        Intensive Care Med. 1996; 22: 707-710https://doi.org/10.1007/bf01709751
        • Knaus W.A.
        • Wagner D.P.
        • Draper E.A.
        • Zimmerman J.E.
        • Bergner M.
        • Bastos P.G.
        • et al.
        The APACHE III prognostic system: risk prediction of hospital mortality for critically III hospitalized adults.
        Chest. 1991; 100: 1619-1636https://doi.org/10.1378/chest.100.6.1619
        • Semeraro F.
        • Ammollo C.T.
        • Caironi P.
        • Masson S.
        • Latini R.
        • Panigada M.
        • et al.
        D-dimer corrected for thrombin and plasmin generation is a strong predictor of mortality in patients with sepsis.
        Blood Transfus. 2019; : 1-8https://doi.org/10.2450/2019.0175-19
        • Bakker J.
        • Nijsten M.W.
        • Jansen T.C.
        Clinical use of lactate monitoring in critically ill patients.
        Ann Intensive Care. 2013; 3: 12https://doi.org/10.1186/2110-5820-3-12
        • Mitra B.
        • Roman C.
        • Charters K.E.
        • O'Reilly G.
        • Gantner D.
        • Cameron P.A
        Lactate, bicarbonate and anion gap for evaluation of patients presenting with sepsis to the emergency department: a prospective cohort study.
        Emerg Med Australas. 2020; 32: 20-24https://doi.org/10.1111/1742-6723.13324
        • Gattinoni L.
        • Vasques F.
        • Camporota L.
        • Meessen J.
        • Romitti F.
        • Pasticci I.
        • et al.
        Understanding lactatemia in human sepsis. Potential impact for early management.
        Am J Respir Crit Care Med. 2019; 200: 582-589https://doi.org/10.1164/rccm.201812-2342OC
        • Shetty A.
        • Sparenberg S.
        • Adams K.
        • Selvedran S.
        • Tang B.
        • Hanna K.
        • et al.
        Arterial to end-tidal carbon dioxide tension difference (CO2 gap) as a prognostic marker for adverse outcomes in emergency department patients presenting with suspected sepsis.
        Emerg Med Australas. 2018; 30: 794-801https://doi.org/10.1111/1742-6723.13095
        • Reynolds R.M.
        • Padfield P.L.
        • Seckl J.R.
        Disorders of sodium balance.
        BMJ. 2006; 332: 702-705https://doi.org/10.1136/bmj.332.7543.702
        • Castello L.M.
        • Baldrighi M.
        • Panizza A.
        • Bartoli E.
        • Avanzi G.C.
        Efficacy and safety of two different tolvaptan doses in the treatment of hyponatremia in the emergency department.
        Intern Emerg Med. 2017; 12: 993-1001https://doi.org/10.1007/s11739-016-1508-5
        • Molaschi M.
        • Ponzetto M.
        • Massaia M.
        • Villa L.
        • Scarafiotti C.
        • Ferrario E.
        Hypernatremic dehydration in the elderly on admission to hospital.
        J Nutr Health Aging. 1997; 1: 156-160
        • Kumar S.
        • Berl T.
        Sodium.
        Lancet. 1998; 352: 220-228https://doi.org/10.1016/S0140-6736(97)12169-9
        • Overgaard-Steensen C.
        • Ring T.
        Clinical review: practical approach to hyponatraemia and hypernatraemia in critically ill patients.
        Crit Care. 2013; 17: 206https://doi.org/10.1186/cc11805
        • Rosner M.H.
        • Ronco C.
        Dysnatremias in the intensive care unit.
        Contrib Nephrol. 2010; 165: 292-298https://doi.org/10.1159/000313769
        • Tsipotis E.
        • Price L.L.
        • Jaber B.L.
        • Madias N.E.
        Hospital-associated hypernatremia spectrum and clinical outcomes in an unselected cohort.
        Am J Med. 2018; 131: 72-82https://doi.org/10.1016/j.amjmed.2017.08.011
        • Polderman K.H.
        • Schreuder W.O.
        • Strack van Schijndel R.J.
        • Thijs L.G.
        Hypernatremia in the intensive care unit: an indicator of quality of care?.
        Crit Care Med. 1999; 27: 1105-1108https://doi.org/10.1097/00003246-199906000-00029
        • Pasantes-Morales H.
        • Cruz-Rangel S.
        Brain volume regulation: osmolytes and aquaporin perspectives.
        Neuroscience. 2010; 168: 871-884https://doi.org/10.1016/j.neuroscience.2009.11.074
        • Scheeren T.W.L.
        • Bakker J.
        • De Backer D.
        • Annane D.
        • Asfar P.
        • Boerma E.C.
        • et al.
        Current use of vasopressors in septic shock.
        Ann Intensive Care. 2019; 9: 20https://doi.org/10.1186/s13613-019-0498-7
        • Beurton A.
        • Teboul J.-.L.
        • Gavelli F.
        • Gonzalez F.A.
        • Girotto V.
        • Galarza L.
        • et al.
        The effects of passive leg raising may be detected by the plethysmographic oxygen saturation signal in critically ill patients.
        Crit Care. 2019; 23: 19https://doi.org/10.1186/s13054-019-2306-z
        • Ni H.-.B.
        • Hu X.-.X.
        • Huang X.-.F.
        • Liu K.-.Q.
        • Yu C.-.B.
        • Wang X.-.M.
        • et al.
        Risk factors and outcomes in patients with hypernatremia and sepsis.
        Am J Med Sci. 2016; 351: 601-605https://doi.org/10.1016/j.amjms.2016.01.027
        • Baldrighi M.
        • Sainaghi P.P.
        • Bellan M.
        • Bartoli E.
        • Castello L.M.
        Hyperglycemic hyperosmolar state: a pragmatic approach to properly manage sodium derangements.
        Curr Diabetes Rev. 2018; 14: 534-541https://doi.org/10.2174/1573399814666180320091451
        • Darmon M.
        • Diconne E.
        • Souweine B.
        • Ruckly S.
        • Adrie C.
        • Azoulay E.
        • et al.
        Prognostic consequences of borderline dysnatremia: pay attention to minimal serum sodium change.
        Crit Care. 2013; 17: R12https://doi.org/10.1186/cc11937
        • Valencia M.
        • Badia J.R.
        • Cavalcanti M.
        • Ferrer M.
        • Agustí C.
        • Angrill J.
        • et al.
        Pneumonia severity index class v patients with community-acquired pneumonia: characteristics, outcomes, and value of severity scores.
        Chest. 2007; 132: 515-522https://doi.org/10.1378/chest.07-0306
        • Tillmann B.
        • Wunsch H.
        Epidemiology and outcomes.
        Crit Care Clin. 2018; 34: 15-27https://doi.org/10.1016/j.ccc.2017.08.001
        • Singer M.
        • Deutschman C.S.
        • Seymour C.W.
        • Shankar-Hari M.
        • Annane D.
        • Bauer M.
        • et al.
        The third international consensus definitions for sepsis and septic shock (Sepsis-3).
        JAMA. 2016; 315: 801-810https://doi.org/10.1001/jama.2016.0287